We are an innovative research lab working to transform the digital future of construction by
...
How can the built environment research benefit from the rapid advancements happening in computer vision? To answer this question, we are investigating the effectiveness of various algorithms (from conventional pattern recognition techniques to emerging deep learning architectures) to enhance the current practice of quality, progress, and safety monitoring of existing facilities, construction sites, and prefabrication shops.
At CCPS we are developing an effective approach for the automated processing of point cloud data (obtained from various sources such as terrestrial laser scanners, RGB-depth sensors, stereo images), to enable 3D reconstruction of the physical objects and automated update of the BIM models accordingly. Robust methods and techniques of computer vision are being investigated to perform various tasks such as 3D object classification, 3D object part segmentation, and semantic segmentation in scenes.
With an integrated access to the contextual and operational data, how new pieces of knowledge can be extracted about the current and future states/behaviors of the built asset? To address this question, we investigate state-of-the-art approaches and techniques of knowledge discovery and soft computing methods, including but not limited to: supervised/ unsupervised machine learning, structure mining, case-based reasoning, expert systems, etc.